A New Biomarker of Fecal Bacteria for Non-Invasive Diagnosis of Colorectal Cancer

BACKGROUND: The intestinal flora is correlated with the occurrence of colorectal cancer. We evaluate a new predictive model for the non-invasive diagnosis of colorectal cancer based on intestinal flora to verify the clinical application prospects of the intestinal flora as a new biomarker in non-invasive screening of colorectal cancer. METHODS: Subjects from two independent Asian cohorts (cohort I, consisting of 206 colorectal cancer and 112 healthy subjects; cohort II, consisting of 67 colorectal cancer and 54 healthy subjects) were included. A probe-based duplex quantitative PCR (qPCR) determination was established for the quantitative determination of candidate bacterial markers. RESULTS: We screened through the gutMEGA database to identify potential non-invasive biomarkers for colorectal cancer, including Prevotella copri (Pc), Gemella morbillorum (Gm), Parvimonas micra (Pm), Cetobacterium somerae (Cs), and Pasteurella stomatis (Ps). A predictive model with good sensitivity and specificity was established as a new diagnostic tool for colorectal cancer. Under the best cutoff value that maximizes the sum of sensitivity and specificity, Gm and Pm had better specificity and sensitivity than other target bacteria. The combined detection model of five kinds of bacteria showed better diagnostic ability than Gm or Pm alone (AUC = 0.861, P < 0.001). These findings were further confirmed in the independent cohort II. Particularly, the combination of bacterial markers and fecal immunochemical test (FIT) improved the diagnostic ability of the five bacteria (sensitivity 67.96%, specificity 89.29%) for patients with colorectal cancer. CONCLUSION: Fecal-based colorectal cancer-related bacteria can be used as new non-invasive diagnostic biomarkers of colorectal cancer. Simultaneously, the molecular biomarkers in fecal samples are similar to FIT, have the applicability in combination with other detection methods, which is expected to improve the sensitivity of diagnosis for colorectal cancer, and have a promising prospect of clinical application.